DerThomy 389bbe48a5 Empty torch cache after optimizer tensor replacement ...

Without the cache emptying I got really high VRAM usage and was not able to train beyond 1 million Gaussians with 24GB VRAM. Now I can train to over 5 million Gaussians.

2024-12-22 13:45:05 +01:00

..

__init__.py

Implementation of 3dgs-mcmc

2024-06-17 12:14:33 -07:00

cameras.py

Provide --data_device option to put data on CPU to save VRAM for training (#14)

2023-07-12 11:30:45 -07:00

colmap_loader.py

Bump rasterizer, faster text loading

2023-09-05 19:30:53 +02:00

dataset_readers.py

Implementation of 3dgs-mcmc

2024-06-17 12:14:33 -07:00

gaussian_model.py

Empty torch cache after optimizer tensor replacement

2024-12-22 13:45:05 +01:00